Predicting Risk of Postoperative Disease Recurrence in Crohn’s Disease: Patients With Indolent Crohn’s Disease Have Distinct Whole Transcriptome Profiles at the Time of First Surgery

Abstract

Background Assessing risk of Crohn’s disease (CD) recurrence following ileocolic resection (ICR) is necessary to optimize medical management and prevent long-term complications. This study aimed to identify noninvasive markers that could predict postoperative disease activity. Methods Inclusion criteria were a diagnosis of CD, first ICR, interval colonoscopy, and whole transcriptome array meeting quality control standards. Demographic and clinical data were obtained from the electronic medical record. RNA extraction and human transcriptome microarray were performed on noninflamed ileal margins from operative specimens. Clinical data and random forest were analyzed in R. Principal components analysis, hierarchical clustering, and pathway enrichment were performed in Partek. Results Sixty-five patients completed the study, and 5 were excluded from analysis due to extreme variability on whole transcriptome analysis. Unsupervised hierarchical clustering revealed that patients with an i0 Rutgeerts score generally segregated from all others. In anti-TNF-naïve patients, unsupervised hierarchical clustering revealed complete segregation of patients with an i0 score. Reduced escalation in therapy and continued mucosal remission, consistent with indolent disease, were seen in the 4 years following surgery. Random forest identified 30 transcripts differentiating i0 patients from the other groups. Pathway enrichment highlighted toll-like receptor, NOD-like receptor, and TNF signaling. This transcriptome signature did not identify i0 anti-TNF-exposed patients. However, anti-TNF-exposed patients with indolent postoperative courses were found to have a transcriptome signature distinct from those with aggressive disease. Conclusions Anti-TNF-naïve and -exposed patients have unique expression profiles at the time of surgery, which may offer predictive value in assessing the risk of nonrecurrence.

Big data in IBD: big progress for clinical practice

“…Cushing et al identified a unique expression profile in anti-TNF-naive and anti-TNF-exposed patients with CD that could predict postoperative disease recurrence. The authors uncovered 30 influential transcripts in anti-TNF-naive patients using random forests-based machine learning models built on demographic and clinical data extracted from the EMR and transcriptomic profile of non-inflamed ileal tissue 103…”

“…The authors uncovered 30 influential transcripts in anti-TNF-naive patients using random forests-based machine learning models built on demographic and clinical data extracted from the EMR and transcriptomic profile of non-inflamed ileal tissue. 103 …”

Clinical applications of artificial intelligence and machine learning‐based methods in inflammatory bowel disease

“…Other studies have focused on the predictive value of genome wide association studies and microbiome data in understanding treatment response. Cushing et al ., for example, identified unique expression profiles in Crohn's disease patients that may predict post‐operative disease recurrence 24 . They looked at gene transcript expression from ileal tissue obtained from operative specimens of Crohn's disease patients to identify unique expression profiles.…”

Artificial Intelligence Enhances Studies on Inflammatory Bowel Disease

“…The ileocolic resection is an essential treatment for a large number of patients with CD after a long disease progression, and the rate of the second or third surgeries gets to 23–43% ( Cushing et al, 2019 ). The prediction of postoperative disease recurrence risk is helpful for clinical decisions and has been proved viable by a study resorting RF algorithm to analyze the whole transcriptome array from 65 patients ( Cushing et al, 2019 ). And 30 influential transcripts in anti-TNF-naive patients uncovered in this study can be the reference for anti-TNF efficacy studies.…”